FWIW: The spacing of the dashes does not seem to be uniform at all.
Chris, there is uniformity there. Lets keep in mind that I only had a few minutes before going to work to record some signals, but I did record on several frequencies in the 1330 to 1350 time period (Dec 19, 2012).
In every recording I made during that time period there is a set pattern.
Pick any single pulse as a starting point. The pulse should be 700 msec long. Now look 11.200 seconds after the start of that pulse, ignoring any other pulses you see prior to 11.200 sec after your selected pulse, there should be another pulse starting exactly 11.200 after your selected pulse. I found this to be the case for my recorded periods, plus and minus my ability to determine start time, about 1 msec. So the period from any pulse to the next pulse from what is probably the same source is 11.200 sec.
Now, divide 11.200 seconds by 700 msec, you get 16. If there are multiple sources and it is some kind of test you would not want to have one pulse stop and the other begin right away, so allow a dead space of 700 msec before the next begins, allowing for 8 sources. However, it looks to me like there is only 4 sources. This means there should be a pulse about each 2.800 seconds. Because of path length distances this will vary.
Call your first selected pulse pulse A0. Call the pulse 11.2 sec after that A1. Call the pulse at 22.4 seconds after the A0 pulse A2, etc, etc.
Roughly 2.8 seconds (this time will be affected by path length differences to you from A0 source and B0 source, in my case this turned out to be 2.672 sec) after pulse A0 you will find another pulse, call this B0. 11.200 seconds after pulse B0 you will find another pulse, call it B01, and so on each 11.200 sec.
Roughly 5.6 seconds after A0 you will find another pulse, again the exact time is dependant on path length (in my case this turned out to be 5.833 seconds), call this pulse C0. Every 11.200 seconds after C0 you will find another “C” pulse.
Roughly 8.4 seconds after A0 you will find another pulse, again the exact time is dependant on path length (in my case this turned out to be 8.055 seconds), call this pulse D0. Every 11.200 seconds after D0 you will find another “D” pulse.
The next pulse after D0 will be A1, and it should be 11.200 seconds after A0.
The gross errors from the 2.8 second spacing, and the fact that one trip around the Earth is only 133.4 msec, would suggest that the system does NOT space the pulses at 2.8 seconds, but rather some value near that. I don’t know, but that was my starting point to look for sets of pulses and a pattern.
All four sources / pulse schedules where seen in my 2 minute recording of 8799 kHz at about 1332 UTC.
T!